Exploratory Analysis of NVSS Multiple Cause-of-Death Mortality Data

Author

Carol Xu

GitHub Repository

Total Overdose Deaths in the U.S. (1999-2020)

Overdose deaths defined by the following ICD-10 Underlying Cause of Death (UCOD) codes:
Code Description
X40–X44 Unintentional/Accidental poisoning by drugs and other biological substances
X60–X64 Suicide/Intentional self-poisoning by drugs and other biological substances
X85 Assault/Homicide by drugs, biological substances, and other substances
Y10–Y14 Undetermined intent, Poisoning by drugs and biological substances

n = 936,279

Top 10 ICD-10 Codes (1999-2020)

See results/top_icd-10_tables for top 20 ICD-10 codes in each year.

Rank Code Description
1 T509 Other and unspecified drugs/medicaments (Acidifying agents, Alkalyzing agents, Immunoglobulin, Immunologicals, Lipotropic drugs, Parathyroid hormones and derivatives)
2 X42 Accidental poisoning by narcotics/psychodysleptics (Incl. Cannabis, Cocaine, Codeine, Heroin, Lysergide, Mescaline, Methadone, Morphine, Opium)
3 X44 Accidental poisoning (unspecified drugs)
4 T404 Other synthetic narcotics (Pethidine)
5 T402 Other opioids (Codeine, Morphine)
6 T405 Cocaine
7 T401 Heroin
8 F191 Mental/behavioral disorders due to psychoactive substance use (Harmful use)
9 T424 Benzodiazepines
10 T436 Psychostimulants with abuse potential

This summary is not a unique number of deaths; Number of occurrences includes duplicates, for death certificates with multiple ICD-10 codes.

Substances of Interest (ICD-10 Codes)

Code Substance
Opioids
T40.0 Opium
T40.1 Heroin
T40.2 Other opioids (morphine, codeine)
T40.3 Methadone
T40.4 Synthetic narcotics (pethidine, fentanyl, etc.)
Stimulants
T40.5 Cocaine
T43.6 Methamphetamine / other psychostimulants
Depressants (non-opioid) / Sedatives
T42.3 Barbiturates
T42.4 Benzodiazepines
Other
T40.7 Cannabis (derivatives)

Opioids

Stimulants

Depressants (non-opioid) / Sedatives

Cannabis

Polysubstance Prevalence

Xylazine (T42.7 & T46.5)

Logistic Regression Predicting Xylazine Involvement (2010-2020)

Xylazine deaths were rare pre 2015, so analyses were done only on 2010-2020 n = 631,387

The model is specified as:

\[ \text{logit}(P(\text{xylazine}=1)) = \beta_0 + \beta_1 \,\text{Year} + \beta_2 \,\text{Male} + \beta_3 \,\text{Age} + \beta_4 \,\text{Black} + \beta_5 \,\text{White} + \beta_6 \,\text{AsianIndian} + \dots + \beta_{10} \,\text{GraduateDegree} + \beta_{11} \,\text{Hispanic} + \beta_{12} \,\text{Married} + \beta_{13} \,\text{Widowed} + \dots \]

Predictor Estimate Std. Error z value Pr(> z
(Intercept) -524.7 14.24 -36.85 <0.001 ***
Year 0.257 0.007 36.47 <0.001 ***
Male (vs Female) -0.284 0.034 -8.33 <0.001 ***
Age (years) 0.0023 0.0014 1.68 0.093 .
Asian Indian (vs Ref) 0.863 0.367 2.35 0.019 *
Black (vs Ref) 0.306 0.186 1.65 0.100 .
Chinese (vs Ref) 0.015 0.607 0.02 0.981
Filipino (vs Ref) 0.142 0.484 0.29 0.769
Guamanian (vs Ref) -10.02 152.3 -0.07 0.948
Hawaiian (vs Ref) -10.13 118.1 -0.09 0.932
Japanese (vs Ref) 1.500 0.424 3.54 <0.001 ***
Korean (vs Ref) 0.389 0.485 0.80 0.422
Other AAPI (vs Ref) 0.497 0.293 1.70 0.089 .
Samoan (vs Ref) -10.14 130.8 -0.08 0.938
Vietnamese (vs Ref) 0.510 0.485 1.05 0.292
White (vs Ref) 0.431 0.181 2.38 0.017 *
Bachelor’s degree 0.233 0.072 3.22 0.001 **
Graduate degree 0.360 0.104 3.46 <0.001 ***
HS / GED 0.053 0.047 1.14 0.255
Some college / Associate 0.062 0.053 1.16 0.244
Education unknown 0.243 0.087 2.78 0.005 **
Hispanic (vs Non-Hispanic) -0.298 0.064 -4.67 <0.001 ***
Married (vs Ref) 0.127 0.049 2.59 0.010 **
Separated (vs Ref) 0.078 0.046 1.70 0.089 .
Unmarried (vs Ref) -0.237 0.149 -1.59 0.112
Widowed (vs Ref) 0.216 0.081 2.66 0.008 **

*** p < 0.001 ** p < 0.01 * p < 0.05 . p < 0.1

Appendix

Data Source

NBER/NVSS Multiple Cause-of-Death Data

  • Multiple Cause-of-death Mortality Data from the National Vital Statistics System (NVSS) of the National Center for Health Statistics (NCHS)
  • National Bureau of Economic Research Public Use Data Archive
  • Downloaded on 4 September 2025